czDrillard,
If you have a use for CPUID and RDTSC, just make sure you change the cpu mode with lines like,
.486
Just keep tweaking it uphill until it builds.
.586 etc ..
CPUID clears the cache so that you reduce the error when using RDTSC.
Regards,
If you have a use for CPUID and RDTSC, just make sure you change the cpu mode with lines like,
.486
Just keep tweaking it uphill until it builds.
.586 etc ..
CPUID clears the cache so that you reduce the error when using RDTSC.
Regards,
I should have posted a warning when using INVD :) he! he! , No problem, Maverick :alright:
CPUID clears the cache so that you reduce the error when using RDTSC.
I've said it many times already, damn I don't want to sound pedantic but I don't think it's pedantic to stress it, since if one needs to clear the cache (and there are very good reasons to need it sometimes) then he will not get that effect by using CPUID, and his/her program will malfunction.
WBINVD = flush the caches in a proper way
CPUID = flush the CPU pipeline
I hope that is useful.
Greets,
Maverick
I'm sure your trying to do something deeper but i found this in the Kernel32.inc
FlushInstrustionCache
FlushInstrustionCache
Sorry Maverick, its brain out of gear, CPUID only clears the pipeline and thats what I have used it for, before RDTSC to stabilise results.
My excuse is senile decay. :tongue:
Regards,
hutch@movsd.com
My excuse is senile decay. :tongue:
Regards,
hutch@movsd.com
Well, I remember I have that too.. thus I can't have it. :tongue:
I mean, I can NOT have Alzheimer, else I would *recall* that I have it. :grin:
Greets,
Mav
I mean, I can NOT have Alzheimer, else I would *recall* that I have it. :grin:
Greets,
Mav